Matrix t-distribution explained

In statistics, the matrix t-distribution (or matrix variate t-distribution) is the generalization of the multivariate t-distribution from vectors to matrices.^{[1] [2]}

The matrix t-distribution shares the same relationship with the multivariate t-distribution that the matrix normal distribution shares with the multivariate normal distribution: If the matrix has only one row, or only one column, the distributions become equivalent to the corresponding (vector-)multivariate distribution. The matrix t-distribution is the compound distribution that results from an infinite mixture of a matrix normal distribution with an inverse Wishart distribution placed over either of its covariance matrices, and the multivariate t-distribution can be generated in a similar way.

In a Bayesian analysis of a multivariate linear regression model based on the matrix normal distribution, the matrix t-distribution is the posterior predictive distribution.

Definition

For a matrix t-distribution, the probability density function at the point

X of an n x p space isf( X ; \nu, M ,\boldsymbol\Sigma, \boldsymbol\Omega) = K x \left| I _{n} + \boldsymbol\Sigma ^{-1} ( X - M )\boldsymbol\Omega ^{-1} ( X - M ) ^{\rm} \right| ^{- (\nu + n + p -1)/(2)} ,

where the constant of integration K is given by

K = (\Gamma _{p\left( (\nu + n + p -1)/(2)} \right))/((\pi) ^{(np)/(2)} \Gamma _{p\left( (\nu + p -1)/(2)} \right)) |\boldsymbol\Omega| ^{- (n)/(2)} |\boldsymbol\Sigma| ^{- (p)/(2)} .

Here

\Gamma _{p} is the multivariate gamma function.

The characteristic function and various other properties can be derived from the generalized matrix t-distribution (see below).

Generalized matrix t-distribution

The generalized matrix t-distribution is a generalization of the matrix t-distribution with two parameters

\alpha and \beta in place of \nu .^[3]

This reduces to the standard matrix t-distribution with

\beta =2 , \alpha = (\nu + p -1)/(2) .

The generalized matrix t-distribution is the compound distribution that results from an infinite mixture of a matrix normal distribution with an inverse multivariate gamma distribution placed over either of its covariance matrices.

Properties

If

X \sim {\rm T} _{n,p} (\alpha,\beta, M ,\boldsymbol\Sigma, \boldsymbol\Omega) thenX ^{\rm} \sim {\rm T} _{p,n} (\alpha,\beta, M ^{\rm} ,\boldsymbol\Omega, \boldsymbol\Sigma).

The property above comes from Sylvester's determinant theorem:

\det\left( I _{n} + (\beta)/(2) \boldsymbol\Sigma ^{-1} ( X - M )\boldsymbol\Omega ^{-1} ( X - M ) ^{\rm} \right) = \det\left( I _{p} + (\beta)/(2) \boldsymbol\Omega ^{-1} ( X ^{\rm} - M ^{\rm} )\boldsymbol\Sigma ^{-1} ( X ^{\rm} - M ^{\rm} ) ^{\rm} \right) .

If

X \sim {\rm T} _{n,p} (\alpha,\beta, M ,\boldsymbol\Sigma, \boldsymbol\Omega) and A (n x n) and B (p x p) are nonsingular matrices thenAXB \sim {\rm T} _{n,p} (\alpha,\beta, AMB , A \boldsymbol\Sigma A ^{\rm} , B ^{\rm} \boldsymbol\Omega B ) .

The characteristic function is^[3]

\phi _{T(Z)} = (\exp({\rm tr)/((i Z ' M ))|\boldsymbol\Omega|) ^{\alpha p} ^{p(\alpha)( 2 \beta)} } |\mathbf'\boldsymbol\Sigma\mathbf|^\alpha B_\alpha\left(\frac\mathbf'\boldsymbol\Sigma\mathbf\boldsymbol\Omega\right),

where

B _{\delta(WZ)} = | W | ^{-\delta} \int _{S>0} \exp\left({\rm tr}( - SW - S ^{-1} Z )\right)| S | ^{- \delta - (12 (p +1 ))/()} d S ,

and where

B _{\delta} is the type-two Bessel function of Herz of a matrix argument.

See also

• Multivariate t-distribution
• Matrix normal distribution

External links

• A C++ library for random matrix generator

Notes and References

1. Zhu, Shenghuo and Kai Yu and Yihong Gong (2007). "Predictive Matrix-Variate t Models." In J. C. Platt, D. Koller, Y. Singer, and S. Roweis, editors, NIPS '07: Advances in Neural Information Processing Systems 20, pages 1721–1728. MIT Press, Cambridge, MA, 2008. The notation is changed a bit in this article for consistency with the matrix normal distribution article.
2. Book: Gupta, Arjun K and Nagar, Daya K. Matrix variate distributions. CRC Press. 1999. Chapter 4.
3. Iranmanesh, Anis, M. Arashi and S. M. M. Tabatabaey (2010). "On Conditional Applications of Matrix Variate Normal Distribution". Iranian Journal of Mathematical Sciences and Informatics, 5:2, pp. 33–43.